Re: [PATCH v2] tmpfs not interleaving properly

[Andrew Morton]

On Thu, 31 May 2012 16:09:15 -0400
KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxx> wrote:

> > --- a/mm/shmem.c
> > +++ b/mm/shmem.c
> > @@ -929,7 +929,7 @@ static struct page *shmem_alloc_page(gfp_t gfp,
> >   	/*
> >   	 * alloc_page_vma() will drop the shared policy reference
> >   	 */
> > -	return alloc_page_vma(gfp,&pvma, 0);
> > +	return alloc_page_vma(gfp,&pvma, info->node_offset<<  PAGE_SHIFT );
> 
> 3rd argument of alloc_page_vma() is an address. This is type error.

Well, it's an unsigned long...

But yes, it is conceptually wrong and *looks* weird.  I think we can
address that by overcoming our peculair aversion to documenting our
code, sigh.  This?

--- a/mm/shmem.c~tmpfs-implement-numa-node-interleaving-fix
+++ a/mm/shmem.c
@@ -927,9 +927,12 @@ static struct page *shmem_alloc_page(gfp
 	pvma.vm_policy = mpol_shared_policy_lookup(&info->policy, index);
 
 	/*
-	 * alloc_page_vma() will drop the shared policy reference
+	 * alloc_page_vma() will drop the shared policy reference.
+	 *
+	 * To avoid allocating all tmpfs pages on node 0, we fake up a virtual
+	 * address based on this file's predetermined preferred node.
 	 */
-	return alloc_page_vma(gfp, &pvma, info->node_offset << PAGE_SHIFT );
+	return alloc_page_vma(gfp, &pvma, info->node_offset << PAGE_SHIFT);
 }
 #else /* !CONFIG_NUMA */
 #ifdef CONFIG_TMPFS
_


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/